Efficiency and productivity of U.S. coal-burning power plants

For the past five decades, statistics from the Energy Information Administration (EIA) reveal an upward trend in electricity demand in the U.S (EIA, 2008, 2009, 2011). The increase in electricity demand encourages electric service providers to implement and maintain up-to-date technologies in their power plants. However, increasing electricity production, ceteris paribus, increases operational cost at power plants. Consequently, power plants constantly aim to ensure electricity demands are met, while keeping cost at a minimum. I use data from 2001 to 2012 on 131 power plants from the EIA, as well as from the Environmental Protection Agency (EPA), the Federal Energy Regulatory Commission (FERC), and the Bureau of Labor Statistics. This paper has two objectives: (i) to uncover how efficient power plants are in minimizing cost and (ii) to determine the factors influencing productivity changes in power plants.

I express the parameters in my model as functions of plant-effects and time-effects. This specification allows me to obtain observation-specific estimates for each parameter. To ensure the estimates conform to economic theory, I apply the Du et al. (2013) constrained weighted bootstrap method. The novelty of this specification is that it allows me to separate overall efficiency into time-invariant and time-varying efficiency measures. Additionally, I decompose productivity into technical change, scale change, input price change, and inefficiency change. The results indicate power plants are quite efficient in minimizing cost. I find evidence of scale economies and technical progress. Consistent with the literature, technical change has the largest effect on productivity.

Inefficiency is costly to power plants because the same level of electricity can be produced at a lower cost. This is interesting from a policy perspective because inefficiency is eventually translated into higher electricity rates for consumers. Although many power plants are in operation for a number of years, there exist productivity differences among them. Quantifying the components that determine these differences could inform public policy aimed at improving productivity.